Using cline for a bit made me realize cursor was doomed. Everything is just a gpt/anthropic wrapper of fancy prompts.
I can do most of what I want with cline, and I've gone back from large changes to just small changes and been moving much quicker. Large refactors/changes start to deviate from what you actually want to accomplish unless you have written a dissertation, and even then they fail.
I agree with all you’ve said but with regards to writing a dissertation for larger changes : have you tried letting it first right a plan for you as markdown (just keep this file uncommitted) and then let it build a checklist of things to do?
I find just referencing this file over and over works wonders and it respects items that were already checked off really well.
I can get a lot done really fast this way in small enough chunks so i know every bit of code and how it works (tweaking manually of course where needed).
But I can blow through some tickets way faster than before this way.
Not if you want custom UI. There are a lot of things you can do in extension land (continue, cline, roocode, kilocode, etc. are good examples) but there are some things you can't.
One thing I would have thought would be really cool to try is to integrate it at the LSP level, and use all that good stuff, but apparently people trying (I think there was a company from .il trying) either went closed or didn't release anything note worthy...
IIRC problem is that VS Code does not allow extensions to create custom UI in the panels areas except for WebViews(?). It makes for not a great experience. Plus Cursor does a lot with background indexing to make their tab completion model really good - more than would be possible with the extensions APIs available.
I use Augment extensively and find it superior to cursor in every way - and operates as an extension. It has a really handy task planning interface and meta prompt refinement feature and the costs are remarkably low. The quality of output implantation is higher IMO and I don’t have to do a lot of model selection and don’t get Max model bill explosions. If there’s something Cursor provided that Augment doesn’t via extension it was not functionally useful enough to notice.
I think Augment has been flying under the radar for many people, and really reserve better marketing.
I've been using Augment for over a year with IntelliJ, and never understood why my colleagues were all raving about Cursor and Windsurf. I gave Cursor a real try, but it wasn't any better, and the value proposition of having to adopt a dedicated IDE wasn't attractive to me.
A plugin to leverage your existing tools makes a lot more sense than an IDE. Or at least until/if AI agents get so smart that you don't need most of the IDE's functionality, which might change what kinds of tooling are needed when you're in the passenger seat rather than the driver's seat.
When the Copilot extension needs a new VS Code feature it gets added, but it isn't available to third party extensions until months later... Err, years later... well, whenever Microsoft feels like it.
So an extension will never be able to compete with Copilot.
Cline and Roo Code (my favorite Cline fork) are fantastic and run as normal VS Code extensions.
Occasionally they lose their connection to the terminal in VSCode, but I’ve got no other integration complaints.
And I really prefer the bring-your-own-key model as opposed to letting the IDE be my middleman.
Using cline for a bit made me realize cursor was doomed. Everything is just a gpt/anthropic wrapper of fancy prompts.
I can do most of what I want with cline, and I've gone back from large changes to just small changes and been moving much quicker. Large refactors/changes start to deviate from what you actually want to accomplish unless you have written a dissertation, and even then they fail.
I agree with all you’ve said but with regards to writing a dissertation for larger changes : have you tried letting it first right a plan for you as markdown (just keep this file uncommitted) and then let it build a checklist of things to do?
I find just referencing this file over and over works wonders and it respects items that were already checked off really well.
I can get a lot done really fast this way in small enough chunks so i know every bit of code and how it works (tweaking manually of course where needed).
But I can blow through some tickets way faster than before this way.
> Wouldn't an extension suffice?
Not if you want custom UI. There are a lot of things you can do in extension land (continue, cline, roocode, kilocode, etc. are good examples) but there are some things you can't.
One thing I would have thought would be really cool to try is to integrate it at the LSP level, and use all that good stuff, but apparently people trying (I think there was a company from .il trying) either went closed or didn't release anything note worthy...
IIRC problem is that VS Code does not allow extensions to create custom UI in the panels areas except for WebViews(?). It makes for not a great experience. Plus Cursor does a lot with background indexing to make their tab completion model really good - more than would be possible with the extensions APIs available.
I use Augment extensively and find it superior to cursor in every way - and operates as an extension. It has a really handy task planning interface and meta prompt refinement feature and the costs are remarkably low. The quality of output implantation is higher IMO and I don’t have to do a lot of model selection and don’t get Max model bill explosions. If there’s something Cursor provided that Augment doesn’t via extension it was not functionally useful enough to notice.
I think Augment has been flying under the radar for many people, and really reserve better marketing.
I've been using Augment for over a year with IntelliJ, and never understood why my colleagues were all raving about Cursor and Windsurf. I gave Cursor a real try, but it wasn't any better, and the value proposition of having to adopt a dedicated IDE wasn't attractive to me.
A plugin to leverage your existing tools makes a lot more sense than an IDE. Or at least until/if AI agents get so smart that you don't need most of the IDE's functionality, which might change what kinds of tooling are needed when you're in the passenger seat rather than the driver's seat.
When the Copilot extension needs a new VS Code feature it gets added, but it isn't available to third party extensions until months later... Err, years later... well, whenever Microsoft feels like it.
So an extension will never be able to compete with Copilot.
As part of this whole drama, the APIs that Copilot uses are being opened up https://code.visualstudio.com/blogs/2025/06/30/openSourceAIE...
It was so they could close source it.
You can ship clised source extensions